Multi-objective concurrent optimization of antenna array parameters for 5G applications

Fifth generation (5G) mobile networks require highly directional beams to serve its users. The basic requirement of beamforming techniques used in 5G is to generate narrow beams with lowest side lobe level (SLL) possible. SLL indicates the concentration of radiated energy in an undesired direction, which leads to power wastage as well as interference with other radiating elements. SLL can be considerably reduced by various methods which include amplitude tapering and different optimization techniques. Tapered excitations reduce the side lobes but increase the beam width. Optimization techniques can be applied to different antenna parameters to achieve reduced SLL without compromising on beam width. Also, these techniques are effective on large arrays. This paper presents and compares elemental excitation optimization and spacing optimization for linear antenna arrays (LAA), using genetic algorithm (GA). Array sizes of 16 and 64 are considered and the excitation amplitudes as well as spacing between elements are varied to achieve reduced SLL. Increase in array size from 16 to 64 resulted in narrower beams as well as reduced SLL, while amplitude and spacing controls improved overall SLL-beam width tradeoff. Spacing control method has performed better than the amplitude-control method in reducing SLL for 16-element and 64-element arrays. Apart from SLL, the half-power beam width (HPBW) is also measured for the optimized arrays, which ensures better concentration of radiated power in desired direction. Compared to the amplitude-control method, HPBW is lowered in spacing control method. To simultaneously reduce SLL and beam width, which are conflicting parameters, multi-objective genetic algorithm (MO-GA) and multi-objective particle swarm optimization (MO-PSO) with amplitude and spacing controls are implemented and the results are compared. Better results are obtained by optimizing both excitation amplitudes and spacing between elements in an array, thus the name concurrent optimization.